How AI and Machine Learning Work: A Lightboard Demonstration of HPE InfoSight

[StorageExperts]

Have you ever had the experience of driving down the road and having your fuel or battery light go on, indicating that your car is running low? Right there you have experienced a recommendation. HPE InfoSight, a cloud-based, AI-driven data analysis tool, provides similar kinds of recommendations for your datacenter. This information is very valuable to IT organizations because, just like you don't want to run out of fuel or electric charge in your car, you don't want to run out of storage space on your arrays.

Simple, straight-line data forecasting features like these can track configuration and performance metrics in your IT infrastructure and keep track of how much space is left on your storage array. When storage is low, HPE InfoSight sends a notification along with an estimate of when you will likely run out of disk space. These estimates are created using very basic rules and parameters. In the strictest sense, this isn't artificial intelligence (AI) or machine learning.

What is Machine Learning?

Machine learning is an iterative process that goes beyond basic rules and thresholds. The process starts with a model for the specific problem it is trying to solve. It takes the first data set, chooses parameters (like thresholds and any other related factors), applies them to the data, and then adjusts the model. It then goes on to look at the next data set, adjusting the model for that set. It does this over and over again. The machine is learning about the problem and how to solve for a set of variables.

In the case of HPE InfoSight, the machine also learns what the best possible solution might be.

When an HPE customer experiences computer system or infrastructure issues, subject matter experts look at the data to determine the potential cause for that problem and give recommendations for a way to resolve it. HPE InfoSight learns this. We call this the AI Recommendation Engine, and it is the most popular part of HPE InfoSight. The AI Recommendation Engine creates a model that can generalize patterns and not memorize them. It is capable of programmatically scanning thousands of sensor array data points from the field of HPE's storage arrays.

To understand what machine learning really is and how the HPE InfoSight AI Recommendation Engine works, a simple example might help. In the following video, I demonstrate how you could teach a machine to tell an almond apart from a grape through a process known as training. Then I go on to show how the AI Recommendation Engine uses data captured from storage arrays in the real world to predict what humans would do when an issue arises.

Watch as I diagram the HPE InfoSight training and recommendation process on a lightboard. How HPE InfoSight leverages AI/ML to deliver an intelligent data platform

Over the past decade, HPE InfoSight has analyzed more than 1,250 trillion data points spanning storage, servers, and virtual machines. Through cloud-based machine learning, it has saved customers over 1.5 million hours by predicting and preventing thousands of disruptions by providing automated recommendations. 

Learn more about HPE InfoSight @ www.hpe.com/infosight

---

Meet Around the Storage Block blogger, Sajjit Thampy. Sajjit is a data scientist who works on HPE Infosight. After 15+ years of experience in the machine learning and data science space with organizations such as Yahoo!, Zynga, and Sun Microsystems, Sajjit is now leading efforts to transform data into actionable insights that enhance HPE’s core offerings.

Sajjit Thampy
HPE InfoSight Data Scientist
Hewlett Packard Enterprise

Storage Experts
Hewlett Packard Enterprise

twitter.com/HPE_Storage
linkedin.com/showcase/hpestorage/
hpe.com/storage